How the Nimble Attainability Index Works: 11 Metrics, Open Data, No MLS Required

Most housing-affordability indices tell you what you already know: housing is expensive. The Nimble Attainability Index asks a more operational question:

Does the housing available in this market align with what local households can actually purchase—and is that alignment getting better or worse?

The current NAI uses 11 components, five families of free public data, and no MLS feed. Every component scores from 0 to 100. Higher means conditions are more attainable for a median-income household.

This is the current v4, attainability-first model.

The 11 components and current weights

| Code | Component | Weight | Primary signal | |---|---|---:|---| | Inc | Income-price gap | 25% | Purchasing power versus home value | | Sup | Supply pressure | 10% | Inventory relative to sales pace | | Mkt | Market health | 10% | Sale-to-list ratio and price cuts | | Mom | Price momentum | 10% | Direction of prices over 12 months | | FHA | FHA headroom | 9% | Distance from the county FHA limit | | Pip | Permit pipeline | 7% | New permits relative to housing stock | | Abs | Absorption | 7% | Share of inventory selling each month | | R/P | Rent-to-price | 7% | Annual rent relative to home value | | F% | FHA utilization | 5% | Home value relative to FHA capacity | | Cmut | Connectivity/commute | 5% | Commute burden | | Eqty | Equity disparity | 5% | Homeownership-rate disparity |

The formula is straightforward:

NAI = Σ(component score × component weight)

The weights sum to 100%. The public calculator uses the same component definitions as the current site model.

Why income-price gap leads the score

The largest weight belongs to the income-price gap because it is the permanent constraint. A loose market helps buyers negotiate, but it does not make an unaffordable home affordable.

The model first translates median household income into a supported purchase price using a housing-payment budget, the current 30-year mortgage rate, and allowances for taxes, insurance, and mortgage insurance. It then compares that purchasing power with the market's home value.

If the home value is at or below the supported price, the component scores 100. As the overshoot widens, the score falls, reaching zero once the home value is 25% above supported purchasing power.

That threshold is intentionally strict. Beyond it, a household generally needs several programs or concessions to line up at once. A market should not receive an “attainable” label merely because inventory is high while the median household remains far short of qualifying.

Static and dynamic signals

Not every metric moves at the same speed.

Slow-moving structural signals include household income, FHA limits, permit pipeline, commute burden, and equity disparity. They describe the constraint landscape.

Fast-moving market signals include inventory, absorption, sale-to-list ratios, price reductions, and price momentum. They describe leverage and direction.

Earlier NAI versions leaned more heavily on dynamic signals because those weights performed better when backtested against near-term buyer outcomes. That improved prediction of whether buyers would gain leverage, but it could also call a city “moderate” while the median household was still far from affording the median home.

V4 makes a deliberate choice: attainability first, market timing second. The score should answer whether households can enter before it answers whether buyers can negotiate.

Where the data comes from

| Source | Inputs | |---|---| | Census American Community Survey | Household income, housing stock, commute burden, homeownership disparity | | Redfin Data Center | Inventory, sales, prices, price reductions, sale-to-list ratio, rents | | Federal Reserve Economic Data (FRED) | Mortgage rates and building-permit series | | U.S. Department of Housing and Urban Development | FHA county loan limits | | Census Building Permits Survey | Residential permit activity |

The current public presentation also uses transparent home-value and source-freshness metadata so visitors can distinguish a current observation from a lagged annual estimate.

What changed from the original nine-metric version

The first public NAI description contained nine core metrics. The current model adds two structural dimensions:

• Connectivity/commute burden, because a lower purchase price can be offset by time and transportation costs.
• Equity disparity, because identical market prices do not produce identical ownership access across households.

The weighting philosophy also changed after backtesting. The earlier model's static variables described current conditions well, while dynamic variables carried more short-horizon predictive power. V4 preserves both, but prevents fast-moving buyer leverage from overwhelming basic purchasing power.

The permit pipeline is now populated from public permit data rather than silently defaulting to a neutral value. That restores a signal the original model intended to measure: whether supply is actually being added.

What the NAI does not measure

The NAI is a market-screening tool, not an underwriting decision.

It does not know a buyer's credit score, savings, debt, employment history, exact mortgage rate, property-tax assessment, HOA dues, or insurance quote. It does not grade school quality or neighborhood safety. County-level and five-year ACS estimates can also blur meaningful differences among cities.

Permit and fee friction matter to what builders can deliver, but they remain separate research layers until comparable evidence and repeated validation justify composite weighting. A precise-looking score should not hide incomplete inputs.

How to use the score

Use NAI to compare markets, identify the size of the purchasing-power gap, and track whether conditions are moving toward or away from attainable housing. Then move from market to parcel, product, and financing assumptions before making an investment decision.

The score is a starting point, not a verdict.

Explore the current NAI calculator and California purchasing-power table.